Multiagent system for joke generation: Humor and emotions combined in human-agent conversation

Abstract

In this paper we present an innovative work on a multiagent joking conversational system. In our research so far we have shown that implementing humor into a chatterbot can visibly improve its performance. The results presented in this paper are the outcome of the next step of our work. They show that a multiagent system, combining a conversational agent, a pun generator and an emotiveness analysis engine, works reasonably well in interactions with users. In the setup used in this research, the emotiveness analysis agent analyses users' utterances and decides whether it is appropriate to tell a pun. Depending on the results of this analysis, the agent chooses either the pun generator, if the decision is that a joke should be told, or the non-humor-equipped agent when the decision is different. Two evaluation experiments were conducted: user (first person) focused and automatic (emotiveness-analysis-based). In both, we compared the performance of the multiagent joking system and a baseline (non-humorous) conversation agent. The results show that in both cases the humor-equipped engine was evaluated as better than the baseline agent. The results are discussed and some ideas for the future are given.

Appropriateness of Ambiguous Facial Expressions Shown by Life-like Agents in Relation to Context of Communication

“…For example, in the field of natural language processing, numerous studies have attempted to estimate emotions and polarity (positive or negative) from conversations among people (e.g. [9,10]). These are a promising key technology.…”

Reducing Excessive Amounts of Data: Multiple Web Queries for Generation of Pun Candidates

“…Also, in one of our earlier works [17] we proposed a humor-equipped chatterbot, able to generate simple puns during nonconstrained conversations with users. The evaluation experiments showed that most users assessed the system with humor as generally better, more interesting, and making them feel better than a similar system without humor.…”

“…Also results of an emotiveness analysis of chat logs showed that in most cases users' emotive states changed to more positive during and after conversations with the system with humor. For more details, see [17].…”

“…Another important issue when implementing the humor generator into chatterbots is the timing of humor, that is, whether jokes are told in appropriate moments of conversations. In our earlier works (e.g., Dybala [17]) we presented an emotiveness-analysis-based timing procedure, in which an emotion-from-text detector was combined with the humorous chatterbot in order to detect users emotions during conversations and on this basis decide whether a joke should be told or not. In the primary setup, the system would tell jokes when users' emotional state was assessed as negative or neutral.…”

“…Evaluation experiments showed that in most cases there was an increase in users positive emotions and decrease in negative emotions during and after the interactions. Also, in a questionnaire conducted afterwards, most users stated that the system used humor in appropriate moments (for more details, see [17]).…”